In standard image coding systems such as ITU-T (International Telecommunication Union Telecommunication Standardization Sector) H.264 and MPEG4 (Moving Picture Experts Group 4)-AVC (Advanced Video Coding), intra-frame prediction (intra prediction) coding has been conventionally employed as Intra-picture prediction coding.
In the intra prediction in ITU-T H.264 and MPEG4-AVC, pixel values of adjacent pixels adjacent to a block to be processed among reconstructed pixel values that are pixel data of blocks adjacent to the block to be processed before a deblocking filter is applied.
Accordingly, predicted pixel values and residual components in each intra prediction mode are added, and one pixel column at the right end and one pixel row at the bottom end of a decoded block are saved as reconstructed pixel values for intra prediction of a next block.
These reconstructed pixel values are saved in an area different from that for decoded pixel values after the deblocking filter is applied thereto, which are pixel data on which decoding is completed.
In the meantime, extension of the numbers of pixels in the horizontal and vertical directions of a macroblock is proposed as an elemental technology for the next-generation image coding standards (see, for example, Non-patent document 1). According to this proposal, it is also proposed that macroblocks of 32×32 pixels and 64×64 pixels, for example, be used in addition to the macroblock size of 16×16 pixels specified in MPEG1, MPEG2, ITU-T H.264 or MPEG4-AVC, for example. In addition, the size of images to be coded and decoded has been recently increasing year after year.